Predicting major donor prospects using machine learning
LE3 .A278 2022
2022
Lee, Greg
Acadia University
Master of Science
Masters
Computer Science
An important concern for many not-for-profits is their major gift fundraising.Major gifts are large gifts (typically $10,000+) and donors who give these gifts are called major donors. Depending upon the charity type, major gifts can constitute as much as 80% of donation dollars received by a charity. Thus, being able to predict who will give a major gift is crucial for charities. In the for-profit sector, using machine learning to target customers has been a long-standing strategy. Charities (i.e., not-for-profits) have been slower to implement machine learning to aid with donor identification and retention. To implement this, we used random forest classifier which predicted with high accuracy (94.47%) and 31 prospects. We tried different experiments using deep learning techniques, Adaboost, decision trees, extra trees classifiers, LASSO, ElasticNet, XGBoost and Light Gradient Boosting regression in order to develop models that can accurately predict future major donors who will donate $10,000 or more. We also experimented with using only donation and behavioural data, which saw increase in false positive values than false negatives for most of the charities. Furthermore, we forecast how much money major donor constituents will contribute to the charity, in which Light GBM regression model was performing well with lowest RMSE values and standard deviations.
The author retains copyright in this thesis. Any substantial copying or any other actions that exceed fair dealing or other exceptions in the Copyright Act require the permission of the author.
https://scholar.acadiau.ca/islandora/object/theses:3730